Studies of waves, turbulence and transport in the Large Plasma Device

Transcription

1 Studies of waves, turbulence and transport in the Large Plasma Device T.A. Carter (+ many others) Dept. of Physics and Astronomy, UCLA

2 Summary/Outline Large Plasma Device (LAPD): Flexible, well-diagnosed experiment for studying basic physical processes (motivations from space, astro, fusion), target for code validation Examples: kinetic/inertial Alfvén waves; fast ion interaction with turbulence, waves; suppression of turbulent transport via sheared flow Control of gradient-driven instability through nonlinear interaction with shear Alfvén waves Uniform plasma: Quasimode driven nonlinearly by beating of co-propagating Alfvén waves In nonuniform plasma, SAW beat wave interacts with gradient-driven instability: unstable mode suppressed (synchronized?) in favor of driven second mode, overall fluctuation amplitude reduced

3 The LArge Plasma Device (LAPD) at UCLA Plasma Source Measurement Region 1 m Cathode Anode 55 cm 17.5 m US DOE/NSF sponsored user facility Solenoidal magnetic field, cathode discharge plasma 0.5 < B < 2 kg,n e cm 3,T e 5 ev,t i 1 ev Large plasma size, D~60cm (1kG: ~300 ρi, ~100 ρs) High repetition rate: 1 Hz Similar parameters to tokamak far edge plasmas: can study basic processes relevant to fusion plasmas (drift turbulence, transport, intermittency, )

4 Example LAPD Users and Research Areas Basic Physics of Plasma Waves: e.g. linear properties of inertial and kinetic Alfvén waves (Kletzing, Howes); Alfvén waves in multi-ion plasmas (Vincena, Maggs, Morales) Physics of fast ion interaction with Alfvén waves and turbulence (Zhou, Zhang, Heidbrink, Carter, Breizman, ) Electron Phase Space Holes (Chen, Kinter, Pickett) Reconnection between flux tubes (Lawrence, Gekelman) Alfvén waves and shocks driven by laser blow-off (Niemann, Gekelman) Nonlinear interactions between Alfvén waves (Brugman, Auerbach, Carter, Vincena, Boldyrev, Howes) Drift-wave turbulence and transport (Pace, Schaffner, Friedman, Popovich, Carter, Maggs, Morales, Horton)

5 Example Profiles CE CE CE Low field case (400G) (also shown: after bias-induced confinement transition); generally get flat core region with D=30-50cm Broadband turbulence generally observed in the edge region

6 Example data: Alfvén wave eigenmodes y (cm) y (cm) Alfvén wave maser (plasma source is resonant cavity) Correlation techniques used to extract structure ~12 hour overnight data run (measurements at ~1600 spatial locations, shots per location) x (cm) J.E. Maggs, G.J. Morales, T.A. Carter, Phys. Plasmas 12, (2005)

7 Dispersion and damping of AWs Recent studies of dispersion/damping for very high k AWs (U. Iowa: Kletzing, Bounds, et al),compared to AstroGK simulation (Howes) kinetic AW inertial AW Kletzing, et al, PRL 104, (2010) Nielson, Howes, et al, PHYSICS OF PLASMAS 17, (2010)

8 Fast ion transport in drift-wave turbulence Limiter used to produce localized, strong drift-wave turbulence Pass Li beam through turbulence, measure beam spreading Coordinated with studies on DIII-D, TORPEX

9 Fast ion transport in drift-wave turbulence Beam spreading observed due to turbulence, increases with lower energy, consistent with orbit-averaging picture [Zhou, et al., submitted]

10 Fast ion/alfvén wave interaction Lithium test ion beam (~1 kev) interacts with antenna launched AWs Doppler-shifted cyclotron resonant interaction observed Interaction only observed with LHP waves Good agreement with theoretical prediction Y. Zhang et al., Phys. Plasmas 16, (2009) Y. Zhang et al., Phys. Plasmas 15, (2008)

11 Discharge Circuit Biased rotation in LAPD + Electrical Breaks B 50Ω Cathode Anode Wall Bias Circuit + 0.2Ω Floating End Mesh Apply positive bias to (floating) wall of chamber (relative to cathode) Radial current in response to applied potential provides torque to spin up plasma Cross-field current carried by ions (through ion-neutral collisions)

12 Dramatic profile steepening observed with biasing (a) (b) (c) CE CE CE As bias exceeds a threshold, confinement transition observed ( H-mode in LAPD) Detailed transport modeling shows that transport is reduced to classical levels during biasing (consistent with Bohm transport prior to rotation) [Maggs, Carter, Taylor, PoP 14, (2007)] [Carter, Maggs, PoP 16, (2009)]

13 Threshold for bias transition is observed, appears to be due to radial flow penetration L n (cm) Bias (Volts) Profile steepening observed for bias above a threshold value

14 Threshold for bias transition is observed, appears to be due to radial flow penetration (a) 8 L n (cm) Bias (Volts) (b) CE Profile steepening observed for bias above a threshold value Flow remains confined to far edge until threshold is exceeded CE

15 Flux probe measurements show suppression of flux at threshold, apparent reversal at larger biases (a) Γ primary electrons present Flux reduced somewhat below threshold, suppressed once flow and shear penetrate to gradient region (b) δi sat As bias is increased, flux reverses direction, apparently indicating inward flux (c) δe θ primary electrons present Electric field fluctuations suppressed more than density, but combination is not enough to explain transport reduction Primary electrons from plasma source appear to affect ability to measure electric field (floating potential)

16 Fluctuation profile modified, but peak amplitude reduced only slightly CE CE CE CE CE CE CE CE CE CE CE CE Broadband drift-alfvén wave fluctuations before biasing Fluctuations concentrate on steepened profile, Doppler shift observed

17 2D Correlation function measurements show dramatic increase in azimuthal correlation, no significant radial decorrelation Unbiased Biased 20cm 10cm Reference is fixed probe 1m away along field No strong evidence for radial decorrelation Decreased transport through longer effective transport step time?

18 Turbulent flux suppression due to modification of densityelectric field fluctuation cross-phase (a) (b) Some amplitude reduction, but not enough to explain transport modification Coherency actually increases with bias (c) (d) Cross-phase term suppressed, reversed with increasing bias, explaining observed behavior of the flux Γ = ñṽ r = ñẽ θ B = 2 Z B Re χ ne (ω)dω = 2 B 0 Z 0 ñ Ẽ θ γ(ω)cos(θ x (ω))dω

19 Modeling LAPD turbulence using BOUT BOUT/BOUT++: 3D Braginskii fluid tokamak edge simulation code modified to simulate LAPD (P. Popovich, M. Umansky, B. Dudson, B. Friedman) Modified to cylindrical geometry, restored neglected terms to handle strong flows in biased LAPD Verification studies: drift waves, KH, rotational interchange γ KH drift waves m Measurements provide challenge to BOUT predictions (validation opportunity). Is Braginskii enough to explain LAPD data? (Gyrokinetic/Gyrofluid simulation of LAPD turbulence??)

20 Nonlinear BOUT LAPD Simulation Simulation uses measured LAPD density profile, but periodic BCs, flows not matched Profile fixed by adding ad hoc plasma source

21 Comparison with measurements: good agreement with spectral shape Power spectrum, [a.u.] LAPD data ν in / Ω ci = ν in / Ω ci = high res. ν in / Ω ci = ν in / Ω ci = f, khz Correlation function from BOUT comparable but longer correlation length predicted Simulation has comparable fluctuation amplitude, intermittency Near term focus: what establishes flow profile in LAPD? Studying Reynolds Stress, Boundary drive (collab with Rogers/Ricci)

22 Alfvén beat wave interactions in LAPD Spontaneous multimode emission by the cavity is often observed, e.g. m=0 and m=1 m=0 m=1 m=0 m=1 T.A. Carter, B. Brugman, et al., PRL 96, (2006)

23 Alfvén beat wave interactions in LAPD Spontaneous multimode emission by the cavity is often observed, e.g. m=0 and m=1 m=0 m=1 Can control multimode emission (e.g. current, shortening the plasma column) m=0 m=1 With two strong primary waves, observe beat driven quasimode which scatters pump waves, generating sidebands Strong interaction: pump δb/b~1%, QM δn/n~10% T.A. Carter, B. Brugman, et al., PRL 96, (2006)

24 KAW beat-wave/instability interaction experiment B 0 = kg Grid Anode Cathode Two independant, perpendicularly polarized Alfvén waves Density Depletion He Plasma Boundary Disk blocks primary electrons Vacuum Chamber Wall Density depletion formed by inserting blocking disk into anode-cathode region, blocking primary electrons therefore limiting plasma production in its shadow Spontaneous growth of instability on periphery of depletion Launch KAWs into depletion, look for interaction Auerbach, et al., arxiv: , PRL accepted

25 Unstable fluctuations observed on depletion (a) m=1 coherent fluctuation observed localized to pressure gradient (d) (e) (b) (c) Drift-Alfvén wave? δn n eδφ k B T e However, Density-potential cross-phase (~180) inconsistent

26 Both pressure gradient and shear flow driven modes unstable (a) (b) (c) Flows/potential gradient also present in density depletion Drift-wave and flutelike (Kelvin-Helmholtz) unstable on measured profiles (linear Braginskii fluid calculation) Nonlinear calculations (BOUT) in progress

27 Resonant drive and mode-selection/suppression of instability instability drift wave beat-driven mode Isat FFT Power (arb, lin) Beat Freq DW Freq Beat Frequency Beat response significantly stronger than uniform plasma case Resonance at (downshifted) instability frequency observed, suppression of the unstable mode observed above (and slightly below) Instability returns at higher beat frequency

28 Suppression of coherent unstable mode and broadband noise BW BW Suppression for beat-driven wave amplitude comparable to unstable mode amplitude With beat wave, quieter at wide range of frequencies (motivates looking at interaction with broadband spectrum)

29 Structure of beat-driven modes 4 No Beat Wave 6 khz Beat Wave 8 khz Beat Wave Phase FFT Power y position (cm) x position (cm) Beat wave has m=1 (6 khz peak), m=2 (8 khz peak) Rotation in electron diamagnetic direction (same as instability)

30 Ring-up/Ring-down observed KAWs on I sat (arb) Average Raw time (ms) kHz wave persists (rings down) after drive is shutoff; ring-up also observed Provides further support for coupling of BW to linear mode (DW/KH)

31 Threshold for control, saturation of BW observed 1.00 I sat FFT Power (arb) 0.10 DW, with 8kHz BW DW, with 0kHz BW 8 khz KAW Antenna Current (A) Modification of DW seen starting at PBW/PDW ~ 10%; maximum suppression for comparable BW power Two effects: electron heating from KAWs modifies profiles, causing some reduction in amplitude without BW BW response seems to saturate as DW power bottoms out

32 Frequency width of control depends on amplitude Window of mode stabilization/control increases with increasing BW drive

33 Parallel BW wavelength matters: weak resonant drive/suppression for counter-propagation 1.0 I sat FFT BW (arb) Co propagating Counter propagating Beat Frequency (khz) Co-propagating BW has small kii, similar to driftwave Counter-propagating mode has short wavelength, expect inefficient coupling to DW/KH (but could couple to IAW ) 20

34 Similar behavior seen using external antenna to excite drift-waves Schroder, Klinger, et al, PRL 86, 5711 (2001) Used external antenna structure on small basic plasma device to try to directly excite drift-waves Saw collapse of spectrum onto coherent drift-wave at the driven frequency (+ harmonics) Using AWs likely more portable to fusion devices? Can we suppress/control instabilities using beat waves driven by, e.g., ICRF?

35 ICRF beat wave control of drift turbulence? ICRF BWs used to excited TAEs in JET [Fasoli, et al.] and ASDEX [Sassenberg, et al.] Explore BW frequency near ITG/TEM/etc mode frequencies, modification of turbulent spectrum, transport? Sassenberg, et al., NF 50, (2010)

36 Summary/Outline Large Plasma Device (LAPD): Flexible, well-diagnosed experiment for studying basic physical processes (motivations from space, astro, fusion), target for code validation Examples: kinetic/inertial Alfvén waves; fast ion interaction with turbulence, waves; suppression of turbulent transport via sheared flow Control of gradient-driven instability through nonlinear interaction with shear Alfvén waves Uniform plasma: Quasimode driven nonlinearly by beating of co-propagating Alfvén waves In nonuniform plasma, SAW beat wave interacts with gradient-driven instability: unstable mode suppressed (synchronized?) in favor of driven second mode, overall fluctuation amplitude reduced